Experimental pitfalls in evaluating vectorial protein secretion in vitro; Sertoli cell secretion of androgen-binding protein and transferrin in two-compartment culture chambers.

We examined the influence of various Millipore filter pretreatments on the amounts of androgen-binding protein (ABP) and transferrin (Trf) found in the outer (OC) and inner (IC) compartment of two-compartment Sertoli cell (Sc) cultures. When Sc were cultured on untreated Millipore filters, less than 10% of ABP was found in OC during 3 initial culture days compared to similar cultures on pretreated filters. Most of the glycoprotein was shown to be bound by the filter. Pretreatment of Millipore filters with 5% bovine serum albumin (BSA) or 2% fetal bovine serum (FBS) maximally saturated the nonspecific protein-binding sites resulting in OC:IC ratio of ABP similar to that found in cultures on polycarbonate membranes, which exhibit very low protein-binding capacity. In contrast to ABP, about 40% of Trf was bound by the Millipore filter on Day 1, with only trace amounts bound thereafter. These differences were due to much higher secretion rate of Trf than ABP, resulting in a relatively smaller fraction of Trf bound to the filter. Again, the nonspecific binding of Trf was greatly reduced by filter pretreatment with 5% BSA or 2% FBS. It is concluded that complete saturation of protein-binding sites of cellulose ester supports is necessary for reliable evaluation of vectorial protein secretion by Sc and other polarized epithelial cells maintained in this type of culture. The implications of partial saturation of protein-binding sites of culture support in interpreting experimental results are discussed.